A global meta-analysis of the effects of land use on the diversity of stream fish and macroinvertebrates

Different catchment land uses affect stream communities through direct and indirect effects by changes in water flow, sediment input, channel physical structure, and productivity. Previous studies of land use effects show a wide variation in the size and direction of biodiversity responses with positive, negative, or even no effects of the loss of native vegetation in catchments. This variation can result from physiographic characteristics, such as the original type of native vegetation (grassland or forest), dominant land use (agricultural or urban), history of change in land use and land coverage (LULC), and climate, and/or from the specific design of each study, such as the range of the analyzed LULC gradient. We performed a meta-analysis of 62 studies using stream macroinvertebrates and fish to determine which of these factors influence the biodiversity responses to changes in LULC. Contrary to our expectation, most of the physiographic factors considered relevant in the literature did not influence the response of macroinvertebrates and fish to land use. We did find, however, that studies with greater ranges of changes in LULC show stronger and the most negative diversity responses for macroinvertebrates. This suggests that studies covering a small gradient in LULC may be unable to detect potentially minute negative impacts on macroinvertebrate diversity. Our results highlight that the response of stream fish and macroinvertebrate diversity to changes in the landscape is context dependent. We suggest that the local environment of study sites and other context-dependent factors should be further investigated to better understand environmental contingencies in stream biodiversity responses.